The interoperation of heterogeneous databases is a pressing need today as organizations attempt to share data stored in legacy databases. These databases are independently developed and maintained to each serve the needs of a single organization. The exchange of data between such databases could be problematic not only because of differences in the representation (syntax) of data but also due to subtle differences in the intended interpretation (semantics) of data. Thus, although translators could be constructed to reformat data from one representation to another, such a translation does not guarantee that the combined, translated data are meaningful – we could be attempting to compare apples with oranges. Therefore, a solution to the interoperation problem must address and resolve both syntactic and semantic differences.

We have developed a query mediation architecture and algorithms for the interoperation of autonomous heterogeneous databases containing data with semantic and representational mismatches. Query mediation in heterogeneous legacy databases makes both the data and the applications accessing the data interoperable. Automated query mediation relieves users from the difficult task of resolving semantic and representational mismatches. Decoupling semantic and representational heterogeneity improves the efficiency of automated query mediation. The approach provides a seamless migration path for legacy databases, enabling organizations to leverage off investments in legacy data and legacy applications.

We have developed a formal policy framework of MAC policies in multilevel relational databases. We have identified the important components of such policies, characterized their desirable properties, and developed sample policies that satisfy these properties. Based on the framework, we compared the MAC policies commonly imposed in or proposed for multilevel relational databases. The framework could be used to capture and resolve the MAC policy mismatches in the secure interoperation of heterogeneous multilevel relational databases. We have also developed techniques for the trusted interoperation of multilevel relational databases that adopt different labeling policies, such as tuple-level and element-level labeling. We have investigated the complexity of trusted interoperation of multilevel relational databases that mismatch in security lattices. We have shown that the detection of security violations through interoperation is PTIME, while the elimination of such violations is NP-complete. Moreover, trusted global interoperation is composable from trusted pairwise interoperation.

Mediated queries can be intolerably inefficient in practice. We are developing implementation-independent techniques for the optimization of multidatabase queries. Cost information is abstracted in a form that is independent of the data models, DBMS implementations, or the hardware platforms of the component databases. The query optimizer decomposes a global query into local queries and generates a global query evaluation plan. We are also developing trusted query mediation techniques for the interoperation of heterogeneous databases, which preserve the autonomy and security of component databases. Queries will be mediated in such a way that neither the rights nor the security of component databases are compromised.

• Dr. Steven Dawson

• Secure Interoperation of Heterogeneous Systems: A Mediator-Based Approach  (1998)
• Query Mediation for Trusted Database Interoperation  (1997)